No routine coagulation laboratory test is recommended during rivaroxaban or dabigatran treatment. However measuring drug concentration and/or anticoagulant activity can be desirable in some special clinical settings, such as bleeding, thrombosis recurrence or emergency surgery. The effects of dabigatran etexilate and rivaroxaban on various coagulation assays have been previously studied in normal plasma spiked with increasing concentrations of the drug. In contrast, few data are available in routinely treated patients. In order to perform and to interpret the results of these tests, it is necessary to determine the usual responses of patient's plasma. We have used several coagulation tests in a prospective study including 106 patients receiving thromboprophylactic treatment with dabigatran 150 or 220 mg od and rivaroxaban 10 mg od for major orthopaedic surgery. The most common tests--prothrombin time (PT) and activated partial thromboplastin time (aPTT)--give results, which vary according to the reagent used. To overcome this limitation, we advocate the use of plasma calibrators, which decreases the inter-laboratory heterogeneity of results. Anti-Xa measurement and Hemoclot, a thrombin diluted clotting assay, are specific assays which have been proposed for rivaroxaban and dabigatran respectively. These tests, conventional PT, aPTT and thrombin generation (TG) have been performed. We demonstrated that measurements of both drugs can determine reliably the drug concentration in patients' plasmas. PT is more prolonged with rivaroxaban than with dabigatran. Interestingly, the pattern of TG was clearly different in relation to the difference in the mechanism of action of the two new anticoagulants. A significant inter-individual variability of response is detected. Rivaroxaban--mean Cmax 140 ng/mL (extremes 0-412) induces a greater increase of PT than dabigatran. aPTT is sensitive to dabigatran. Rivaroxaban concentrations were in good agreement with two other studies while unexplained lower than expected concentrations were found in dabigatran patients receiving 220 mg once a day [mean Cmax 60 ng/mL (extremes 0-320)]. An interference by pantoprazole, a drug which reduces dabigatran absorption, could explain the observed lower than expected results.